Generative AI companies are facing global accusations of widespread copyright infringement for training models on protected works without permission [1, 2].
This conflict represents a fundamental clash between the rapid scaling of artificial intelligence and the legal protections afforded to creators. As AI companies ingest vast amounts of data to improve their models, artists, musicians, and writers argue that their intellectual property is being stolen to build commercial products that may eventually replace them [4, 5].
In the music industry, the scale of the issue is immense. The International Confederation of Music Publishers (ICMP) said "the totality of the music mondiale is aspirated by the AI giants" [3], suggesting that 100% of the world's music has been utilized by these systems [3].
Legal responses are emerging globally. In France, a law was adopted on April 8, 2026 [1], which reverses the burden of proof, requiring AI providers to prove they had the right to use the data they trained on.
However, the legal landscape remains contradictory. While some experts argue that these models violate copyright on a global scale [1], a preliminary judgment in the U.S. case of Bartz v. Anthropic suggested that the use of legally acquired works for AI training may be permitted [6].
Some critics argue that these developments could signal the end of copyright as a viable legal concept [5]. To prevent further damage, some global authorities have called for a six-month moratorium on the development of certain AI systems to allow legal frameworks to catch up [5].
Cécile Rap-Veber of the SACEM has been among those highlighting the need for stricter licensing requirements to ensure creators are compensated when their work fuels AI growth [1, 7].
“"the totality of the music mondiale is aspirated by the AI giants"”
The tension between AI training and copyright law is shifting from theoretical debate to active litigation and legislation. By reversing the burden of proof in France and testing 'fair use' boundaries in the U.S., the global community is attempting to determine if AI training constitutes a transformative new use or a massive scale of unauthorized reproduction. The outcome will likely dictate the future economic model for all creative industries.
